ML Classification Methods Comparison for Breast Cancer Diagnosis in Clinical Application Field

Abstract

With the development of computing resources, deep learning method has made great progress in the field of image recognition. And as a result, deep learning-based image detection methods have been able to make great breakthroughs in medical fields such as disease recognition, microbiological testing, bacterial quantification, and tumour detection. With the advancement of in the ML field, the computer-aid diagnosis system has been improving. This paper reviews basic CAD ML system, and the recently improved DL system. In general, both methods offer the ability to assist the medical expert in giving them opinions. Deep learning, on the other hand, shows promising improvement in breast cancer diagnosis, but still with some limitations, such as requirement of high computational power and cost, as well as data scarcity. The accuracy rate of DL can reach up to 97%, which is the best in the field. Thus, both effectively reduce false positive rate of detecting breast cancer, improving the performance for detection and diagnosis.